Derlin-1 promotes ubiquitylation and degradation of the epithelial Na⁺ channel, ENaC Free

Hypertension is a major risk factor for heart attack, stroke and kidney failure. The regulation of blood pressure is an important function of the kidney, which maintains control of features such as Na⁺ balance and blood volume (Ayuzawa and Fujita, 2015). Na⁺ reabsorption and fluid balance are maintained by the kidney through the action of apical membrane Na⁺ channels in the distal nephron, making these channels important contributors to the development of hypertension (Rossier, 2014).

One of these apical membrane Na⁺ channels, the epithelial Na⁺ channel (ENaC), is a major target for pathways that elicit fine control over distal Na⁺ reabsorption. The ENaC localizes to the apical membranes of principal cells in the aldosterone-sensitive distal nephron (ASDN), where it functions in the regulation of whole-body Na⁺ homeostasis and the control of extracellular fluid volume and blood pressure. This function is demonstrated most clearly by genetic disorders of Na⁺ channel activity (e.g. Liddle Syndrome and pseudohypoaldosteronism type 1), which produce significant and opposing effects on blood pressure (Kashlan and Kleyman, 2011).

The ENaC is comprised of three homologous subunits, α, β and γ (hereafter α-ENaC, β-ENaC and γ-ENaC; also known as SCNN1A SCNN1B and SCNN1G); each subunit contains two transmembrane domains, an extracellular loop, and intracellular N- and C-termini (Butterworth, 2010). Most (75%) of these subunits are initially localized in the endoplasmic reticulum (ER) lumen and membrane, while the remaining subunits (25%) are located in the cytoplasm (Snyder et al., 1994). Upon assembly in the ER, the ENaC is trafficked to the plasma membrane, where it functions as a channel. The expression of α-ENaC takes priority and is higher than the expression of the other two subunits (Snyder, 2005). The α-ENaC expression serves to stabilize β-ENaC and γ-ENaC, thereby favoring ENaC assembly (Masilamani et al., 1999). In the absence of α-ENaC, the β-ENaC and γ-ENaC subunits are degraded by the proteasome (Snyder, 2005; Staub et al., 1997; Valentijn et al., 1998). However, unlike the case for their well-investigated lysosomal degradation pathways, the underlying mechanism that leads to proteasome-mediated degradation of ENaC subunits remains unclear (Ronzaud and Staub, 2014; Rotin and Staub, 2011).

ER-associated protein degradation (ERAD) is a process in which protein substrates in the ER are polyubiquitylated and moved into the cytosol for proteasomal degradation (Römisch, 2005). ERAD can be divided into four steps: recognition, retrotranslocation, ubiquitylation and degradation. Protein substrates are initially recognized and then removed from the ER; this is followed by their cytoplasmic ubiquitylation and the ubiquitin-dependent proteolysis by the 26S proteasome (Bagola et al., 2011).

One function of ERAD is to eliminate misfolded proteins, which is carried out by different mechanisms. This elimination is best characterized in yeast systems, where distinct E3 complexes define different ERAD pathways (Carvalho et al., 2006; Vashist and Ng, 2004). Proteins destined for proteasomal degradation, such as membrane and soluble proteins with luminal lesions, are delivered to the ERAD-L pathway. There, the Hrdlp or Hrd3p ligases form a complex by binding to Derlp via the linker protein Usalp (Denic et al., 2006). Membrane proteins with misfolded cytoplasmic domains use the ERAD-C pathway and are directly targeted to the DoalOp E3 ligase. Substrates with misfolded intramembrane domains are directed to the ERAD-M pathway, which differs from ERAD-L by virtue of its independence from Usalp and Derlp (Carvalho et al., 2006).

The ERAD of the ENaC has also been studied in yeast. In that pathway, the Hsp40 proteins can function independently of a cognate Hsp70 to facilitate the degradation of ENaC subunits, and the turnover of each ENaC subunit requires the E3 ubiquitin ligases required for both the ERAD-L/M and ERAD-C pathways (Buck et al., 2010). However, the process controlling the recognition of ENaC subunits by the ERAD machinery remains to be established in mammalian cells.

Many studies have demonstrated that derlin-1 forms a retrotranslocation channel in the ER membrane, and that it may also function in the recognition of misfolded membrane or transmembrane proteins. Derlin-1 interacts with recognized target proteins of ubiquitin-mediated proteasomal degradation in the cytosol (Lilley and Ploegh, 2004; Ye et al., 2004). Retrotranslocated proteins are ubiquitylated by the E1–E2–E3 ubiquitin system, and the corresponding ubiquitin-conjugating enzyme (E2) and ligase (E3) are specific for the three different ERAD pathways (Carvalho et al., 2006; Denic et al., 2006). Once a substrate is ubiquitylated, its proteasomal degradation takes place in the cytosol.

Sun et al. and other groups have investigated the physiological role of derlin-1 on the biogenesis of the cystic fibrosis transmembrane conductance regulator (CFTR) (Sun et al., 2006), and have demonstrated the involvement of derlin-1 in the degradation of CFTR as a mammalian ERAD substrate (Wang et al., 2008; Younger et al., 2006). This raises the possibility of an involvement of derlin-1 in ENaC recognition, ubiquitylation and degradation in mammalian cells. The goal of the present study was therefore to determine the mechanism by which ENaC ubiquitylation and degradation takes place.

The physiological significance of derlin-1 in the regulation of endogenous ENaC expression was evaluated in mouse cortical collecting duct (mCCD) cells transfected with or without pcDNA3.1-derlin-1, or in knockdown experiments performed with shRNA targeting derlin-1 expression, and the results were compared with those obtained with a control scrambled shRNA. As shown in Fig. 1A, the endogenous expression of α-ENaC, β-ENaC, and γ-ENaC were significantly decreased by overexpression of derlin-1 in mCCD cells. Conversely, a dramatic upregulation in the levels of the three ENaC subunits was observed with endogenous derlin-1 knockdown in mCCD cells. In addition, we examined the possibility that derlin-1 regulates the expression of ENaC by overexpressing α-ENaC, β-ENaC and γ-ENaC, with or without derlin-1 transfection, in HEK293 cells and detecting expression by western blotting. As shown in Fig. 1B, derlin-1 overexpression markedly decreased the expression of all three ENaC subunits at the protein level. The mean data for expression of the three ENaC subunits are shown in Fig. 1C. We confirmed that derlin-1 affects ENaC subunit expression at the transcript level by using real-time quantitative PCR (qPCR) to determine the expression of mRNA for the three ENaC subunits after co-transfection. As shown in Fig. 1D, none of the mRNA levels showed significant differences in the presence or absence of derlin-1 in HEK293 cells. These findings verified that derlin-1 suppresses ENaC expression at the post-translational level.

We next examined the interaction of derlin-1 with α-ENaC in co-immunoprecipitation (co-IP) experiments. Derlin-1 antibodies were used to isolate protein complexes from mCCD cells, and these complexes were subsequently blotted with anti-α-ENaC antibodies. As shown in Fig. 2A, IPs performed with IgG as a control yielded no α-ENaC signal, and the physical interaction between derlin-1 and α-ENaC was detected in the co-IP analysis of mCCD cells. We also confirmed the binding signals of α-ENaC with derlin-1 in IPs performed using anti-α-ENaC antibody and subsequent blotting with derlin-1 antibodies (Fig. 2B). To further confirm derlin-1 binding with α-ENaC, we examined the interaction between α-ENaC and derlin-1 in HEK293 cells that had been transfected with α-ENaC and HA–derlin-1. Anti-HA antibodies were used to isolate protein complexes from HEK293 cells, as shown in Fig. 2C, HA–derlin-1 was co-immunoprecipitated with α-ENaC, but not with IgG in the control group. The signals for binding between α-ENaC and HA–derlin-1 were also detected in HEK293 cells by IPs performed by using anti-α-ENaC antibodies (Fig. 2D), suggesting a specific interaction between these two proteins. Importantly, the double immunofluorescence study indicated that α-ENaC and derlin-1 were partially colocalized in mCCD cells (Fig. 2E) and HEK293 cells (Fig. 2F), in agreement with the biochemical data depicted in Fig. 2A–D.

We investigated the α-ENaC-specific binding region of derlin-1, as shown in Fig. 3A, by constructing two HA-tagged derlin-1 deletion mutants: HA–dNt (deletion of the N-terminal 17 amino acids) and HA–dCt (deletion of the C-terminal 74 amino acids). Co-IPs were performed to determine the interaction between the expressed α-ENaC and derlin-1 proteins. Consistent with Fig. 2, binding occurred between α-ENaC and wild-type (WT) derlin-1. The α-ENaC protein also showed a marked interaction with derlin-1 dNt, as well as with derlin-1 dCt, when IPs were performed using anti-HA antibodies (Fig. 3B). Positive α-ENaC-binding signals were also detected for derlin-1 dNt and derlin-1 dCt using anti-α-ENaC antibodies for the IP (Fig. 3C). These results suggest that the functional domain associated with α-ENaC is part of the membrane-anchored domains or the loop regions of derlin-1, rather than the N- or C-terminals. IPs performed with an anti-IgG as a control yielded no HA–derlin1 signal (Fig. S1).

We investigated the mechanism of derlin-1 regulation of α-ENaC protein expression by transient transfection of HEK293 cells with or without HA–derlin-1 plasmids. The day after transfection, the cells were treated with cycloheximide and Brefeldin A for 0, 2, 4 or 6 h to block both the synthesis and maturation of new polypeptides. We then examined α-ENaC protein expression by western blotting. As shown in Fig. 4A, α-ENaC disappeared at a much faster rate from cells overexpressing derlin-1. After a 2 h chase, only 50% of the initial amount of α-ENaC remained, whereas 75% was still present in the control condition (Fig. 4B). Therefore, derlin-1 overexpression in HEK293 cells appeared to induce a shorter half-life for α-ENaC protein expression. The findings suggest that derlin-1 physically interacts with α-ENaC and might function in ENaC degradation.

As shown in Fig. 5A, treatment with MG132, a proteasome inhibitor, increased the expression of α-ENaC in mCCD cells. The mean data for α-ENaC expression, shown in Fig. 5B, indicated that the regulation of α-ENaC might occur via the ubiquitin–proteasome degradation pathway. Transient infection of the HEK293 cells with HA-tagged ubiquitin (HA–Ub) and α-ENaC constructs resulted in a greater expression of ubiquitin in the cells overexpressing derlin-1 than in the cells that were not transfected with derlin-1 (Fig. 5C). The mean data for ubiquitin expression are shown in Fig. 5D. Conversely, downregulation of derlin-1 in HEK293 cells by transfection with derlin-1 shRNA significantly decreased ubiquitin expression (Fig. 5E); the quantified data are shown in Fig. 5F. derlin-1 expression also promoted the ubiquitylation of α-ENaC; more than twice as much ubiquitin was conjugated to α-ENaC in HEK293 cells transfected with HA–derlin-1 than in cells transfected with vector only (Fig. 5G). The quantified data are shown in Fig. 5H. These findings suggest that derlin-1 expression promotes the ubiquitylation and degradation of α-ENaC.

We sought a better understanding of ubiquitin-mediated α-ENaC degradation and to eliminate indirect effects of proteasome inhibitors on α-ENaC degradation by employing a ubiquitin construct in which all seven lysine residues were mutated to arginine residues (denoted KO). This mutation abolishes the ability of ubiquitin to form a polyubiquitin chain (polyUb) on the target proteins (Lim et al., 2005). We also made a HA-tagged ubiquitin containing lysine at specific positions only (the rest were mutated to arginine residues, K-to-R): position 6 (K6), 11 (K11), 27 (K27), 29 (K29), 33 (K33), 48 (K48), or 63 (K63). The expression of the polyUb chain and the extent of ubiquitin-mediated α-ENaC degradation were then examined by co-transfection of HEK293 cells with α-ENaC and these ubiquitin mutants. The KO mutant is unable to form a polyUb chain, so this mutant was used as a standard control. As shown in Fig. 6A, expression of the polyUb chain was markedly higher after cells were transfected with lysine linkage mutant K11 than with KO or the other mutants. The quantified data are shown in Fig. 6B. Further confirmation of the requirement for K11 for polyUb chain linkage to α-ENaC was obtained by generating single lysine residue mutations to arginine (K to R), including K11R and K48R. When compared to the wild-type HA–Ub, the expression of KO or K11R significantly decreased polyUb chain linkage to α-ENaC, while K48R expression had no effect on that linkage (Fig. 6C,D).

The possibility that the K11-linked polyUb chain would facilitate α-ENaC degradation was tested by transfecting HEK293 with derlin-1 and the different ubiquitin mutants, followed by immunofluorescence staining to examine α-ENaC expression. Both HA–Ub WT and K48R were clearly effective at decreasing α-ENaC expression; however, the KO and K11R mutations had no effect on α-ENaC expression because they prevented the formation of a polyUb chain that would facilitate α-ENaC degradation (Fig. 6E,F), in agreement with the biochemical data about α-ENaC expression in the whole-cell lysates depicted in Fig. 6C. Therefore, these novel findings suggest that the K11-linked polyUb chain is the most effective at facilitating α-ENaC degradation.

Next, we used liquid chromatography and tandem mass spectrometry (LC-MS/MS) to identify the specific ubiquitin E3 ligases that assemble polyUb chains with derlin-1. The cell lysates were immunoprecipitated with anti-derlin-1 antibody, gels were stained (Fig. 7A) and the samples were analyzed by LC-MS/MS. The peptides identified by LC-MS/MS were categorized based on their functions in protein posttranslational regulation, such as ubiquitylation, phosphorylation, methylation, glycosylation, acetylation and SUMOylation. As shown in Table S1, HUWE1 was one of the candidate substrates for derlin-1 that showed an interaction related to ubiquitylation. HUWE1 expression was clearly induced, but α-ENaC expression was decreased by derlin-1 transfection of HEK293 cells (Fig. 7B). The mean data for HUWE1 expression are shown in Fig. 7C. The possibility of binding between HUWE1 and α-ENaC was further tested by performing a co-IP experiment in mCCD cells. The data supported a physical interaction between α-ENaC and HUWE1 (Fig. 7D).

To evaluate that HUWE1 is indeed the E3 ligase for α-ENaC, we regulated HUWE1 expression in mCCD cells by transfection with or without pcDNA3.1-HUWE1, or we knocked HUWE1 expression by transfecting siRNA targeting HUWE1 in mCCD cells. As shown in Fig. 7E, HUWE1 expression was significantly increased by 2-fold upon pcDNA3.1-HUWE1 transfection, and successfully decreased by ∼60% by siRNA-HUWE1 knockdown in mCCD cells. We found that HUWE1 over-expression significantly decreased the α-ENaC level, and knockdown of HUWE1 dramatically suppressed the degradation of α-ENaC modulated by derlin-1 (Fig. 7F). These data indicate that HUWE1 is the E3 ubiquitin ligase that functions in derlin-1 regulation of ubiquitin-mediated α-ENaC degradation.

We demonstrated in this study that derlin-1 recognized α-ENaC and promoted its ubiquitin-mediated degradation in mammalian cells. derlin-1 physically interacted with α-ENaC and reduced its expression. Knockdown experiments using derlin-1 shRNA demonstrated that decreased expression of derlin-1 markedly decreased the total ubiquitin level, but increased the expression of α-ENaC. These findings implied that endogenous derlin-1 was involved in the recognition and ubiquitin-mediated degradation of α-ENaC. The main function of ENaC in the kidney distal nephron are the mediation of transmembrane Na⁺ transport, regulation of Na⁺ and water balances, and stabilization of blood pressure. Dysfunction of ENaC can therefore cause blood pressure dysregulation.

ENaC expression and transport depends on ENaC synthesis and degradation. The lysosome-mediated degradation pathways of ENaC have been well investigated (Snyder, 2005), but significantly less research has focused on the molecular mechanism underlying proteasome-mediated degradation of ENaC subunits. Recently, Buck et al. reported several studies confirming the ERAD of ENaC and related mechanisms in a yeast system (Buck et al., 2010). They discovered that Jem1 and Scj1 assist in ENaC ubiquitylation, and that overexpression of ERdj3 and ERdj4 (also known as DNAJB11 and DNAJB9, respectively), two luminal mammalian Hsp40 proteins, increased the proteasome-mediated degradation of ENaC in vertebrate cells. Their data indicated that Hsp40 proteins, independently of Hsp70, might select substrates for ERAD, and that ENaC required unique molecular chaperones for ERAD. They also determined that the chaperone Lhs1/GRP170 (also known as HYOU1) selects the nonglycosylated form of the α-ENaC for ERAD (Buck et al., 2013). Our studies on the ENaC regulation by derlin-1 provide more evidence for ENaC as a substrate for an ERAD quality control system and contribute additional information regarding the pathways of ENaC degradation.

Derlin-1 belongs to the derlin family and promotes protein degradation as an E3 ubiquitin ligase modulator. Three derlin isoforms are found in the human genome, derlin-1, -2, and -3. derlin-2 and -3 share ∼75% identity, whereas derlin-1 and -2 or derlin-1 and -3 share only ∼30% identity. All three isoforms are localized in the ER membrane (Oda et al., 2006). Derlin-1 contains four transmembrane segments, with both the N- and the C-termini located in the cytosol. We investigated domain-specific mediated ENaC degradation by constructing derlin-1 deletion mutants and co-expressed them with WT α-ENaC in HEK293 cells. Deletion of the N-terminus (dNT) or the C-terminus (dCT) did not change the binding signals between derlin-1 and α-ENaC, indicating that the membrane-anchored domains or the loop regions of derlin-1 are crucial for α-ENaC degradation. Further studies will be needed to identify the other potential motif(s) located within these regions of derlin-1 that operate in the ENaC degradation pathway.

The substrates of derlin-1 identified by LC-MS/MS analysis displayed diverse protein modification functions. Most of the substrates were ubiquitin-related proteins, while others participated in protein phosphorylation, methylation, glycosylation, acetylation and SUMOylation. These protein modifications might be also related to ENaC, apart from being the substrates of derlin-1, because the cell lysates of the MS sample were from HEK293 cells that overexpressed α-ENaC. Many more studies are needed to determine the details of the molecular mechanism underlying ENaC regulation by derlin-1.

The role of polyUb chains in ENaC ubiquitylation was explored by using diverse mutants of all seven lysine residues in ubiquitin. Our results provide the first indication that the K11-linked polyUb chain is the most effective at facilitating α-ENaC degradation. This novel finding will allow us to identify different ENaC biogenesis pathways, because different lysine linkages of polyUb chains perform different cellular functions.

Ubiquitylation requires a multi-step, enzymatic, thiol-mediated transfer process involving a ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzymes (E2) and ubiquitin ligases (E3). This process begins with activation of the C-terminus of ubiquitin by an E1, transfer of ubiquitin to the conserved active site cysteine of an E2 through a transesterification reaction, and the relay of ubiquitin to a substrate via an E3 (Rotin and Staub, 2011). The substrate specificity for ubiquitin transfer to a target protein is conferred by the ubiquitin E3 ligases (Hurley and Stenmark, 2011). Nedd4-2 (also known as NEDD4L) is an E3 ubiquitin ligase with a known involvement in the degradation of ENaC (Rotin and Staub, 2011). The yeast system requires Hrd1 and Doa10, two E3 ubiquitin ligases, to present a critical level of ubiquitin for effective targeting of ENaC for proteasome-mediated degradation, because the degradation and ubiquitylation of the ENaC subunits is suppressed by deletion of the genes encoding Hrd1 and Doa10 (Buck et al., 2010). Fig. 7 in the present study identified and confirmed one E3 ubiquitin ligase, HUWE1, which was physically associated with α-ENaC as well as derlin-1 in mCCD cells; we found that HUWE1, which contains a C-terminal HECT (an E6AP-type E3 ubiquitin protein ligase) domain, contributed to ubiquitin-mediated α-ENaC degradation together with derlin-1.

In summary, as shown in Fig. 8, derlin-1 interacts physically with α-ENaC to initiate α-ENaC retrotranslocation. In addition, HUWE1, an ER-resident E3 ubiquitin ligase, is recruited and assists in the transfer of a K11-linked polyUb chain to α-ENaC, leading to the formation of an α-ENaC ubiquitin-mediated degradation complex. These findings suggest that the derlin-l pathway represents a significant early checkpoint in the recognition and degradation of ENaC in mammalian cells.

Antibodies against derlin-1 and β-actin were purchased from Sigma-Aldrich (St Louis, MO). The antibodies specific for ubiquitin, tubulin and HA-tag were purchased from Cell Signaling Technology (Danvers, UK), Bioworld (Minneapolis, MN) and Abmart (Shanghai, China), respectively. The antibodies against α-ENaC, β-ENaC and γ-ENaC were obtained from StressMarq Biosciences (YYJ, CN). A polyclonal antibody against α-ENaC, as described previously (Liang et al., 2006), was used in some experiments in this study. The antibody against HUWE1 was purchased from Proteintech (Chicago, USA). Other antibodies including normal rabbit IgG for immunoprecipitation and secondary mouse or rabbit antibodies for western blotting were purchased from Thermo Fisher (Waltham, UK). The catalog numbers and dilutions of all antibodies are shown in Table S2. MG132 was from Enzo Life Science (NY). Cycloheximide and Protein A–agarose beads were purchased from Sigma-Aldrich. The scrambled siRNA and siRNA against mouse HUWE1 were commercially obtained as SMARTpool^® reagents from GE Dharmacon Inc. The constructs pcDNA3.1-Derlin1 and pcDNA3.1-HUWE1 were gifts from Dongming Su (Nanjing Medical University, Nanjing, China). α-ENaC, β-ENaC and γ-ENaC were amplified from an ultimate open reading frame clone by PCR, and cloned into pcDNA3.1. The constructs of wild-type HA-tagged derlin-1, shRNA-derlin-1, shRNA-scramble, N-terminal deletion HA-tagged derlin-1 (dNt) and C-terminate deletion HA-tagged derlin-1 (dCt) were designed and cloned into pcDNA3.1. The mutations of ubiquitin lysine residues were generated using the site-directed mutagenesis kit (Stratagene). Wild-type ubiquitin and the mutated ubiquitin plasmids were amplified by PCR and cloned into pcDNA3.1 sequentially. All plasmid constructs were confirmed by DNA sequencing.

HEK293 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C and in 5% CO₂. The mCCD cells (provided by Alain Vandewalle and Marcelle Bens, INSERM, Paris, France) were grown in flasks (passage 30–40) in mixed medium as previously described (Liang et al., 2010; Shen et al., 2015), namely equal volumes of DMEM and F12 containing 5 mg/ml transferrin, 60 nM Na⁺ selenate, 50 nM dexamethasone, 2 mM glutamine, 1 nM triiodothyronine, 10 ng/ml epidermal growth factor, 5 μg/ml insulin, 20 mM D-glucose, 2% FBS and 20 mM HEPES pH 7.4, at 37°C and 5% CO₂.

For cell transfection, the constructed vectors were transfected into HEK293 cells or mCCD cells using X-treme GENE HP (Roche). At 4–6 h after transfection, the medium was complemented blood serum, and the cells were maintained in an incubator at 37°C and 5% CO₂. Gene expression or knockdown was examined by western blotting at 24–48 h after gene transfection.

Total RNA was extracted from HEK293 cells by Trizol reagent (Invitrogen) according to the manufacturer's instructions. Reverse transcription of total RNA was performed by a commercial kit (Takara). A 10 μl mixture containing 1000 ng RNA and a set of gene-specific primers were subjected to real-time PCR using ABI Plus One Step (BioRad). The sequences of the primers used in the qPCR were as follows: α-ENaC, forward 5′-TCTGCACCTTTGGCATGATGT-3′ and reverse, 5′-GAAGACGAGCTTGTCCGAGT-3′; β-ENaC, forward, 5′-AGACAACCACAATGGCTTAACA-3′ and reverse, 5′-TGAGGCTACATAGTCTCATGGC-3′; γ-ENaC, forward, 5′-GCACCCGGAGAGAAGATCAAA-3′ and reverse, 5′-TACCACCGCATCAGCTCTTTA-3′; GAPDH, forward, 5′-CCCCTTCATTGACCTTCAACTA-3′ and reverse, 5′-GAGTCCTTACGATACCAAAG-3′.

The lysates from HEK293 cells or mCCD cells were resolved by 8% or 12% SDS-PAGE on gels that contain 50 mM Tris-HCl, pH 6.8, pH 8.8, 10% ammonium persulfate, 30% Acr-Bis, 10% SDS and 1% TEMED. Unbound sites were blocked with 5% nonfat milk in Tris-buffered saline (TBS) with 0.1% Tween-20 for 1.5 h at room temperature. The blotted membranes were probed with the indicated primary antibodies at various dilutions, and then incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies. The immuno-recognition signals were detected using enhanced chemiluminescence (ECL) and were acquired using an Image Quant ECL system (PerkinElmer Life Sciences, Wellesley, MA). Western blotting data were quantified with Image Lab software.

For IP, lysates were collected from HEK293 or mCCD cells after transient transfection or drug stimulation, and lysed with NP40 that included 10 mM Tris-HCl pH 7.4, 10 mM NaCl. The concentrations of the lysates were examined by protein assays (BCA), and then 0.5 mg protein was incubated with specific moderate antibodies overnight at 4°C. 100 μl protein A beads were incubated with the protein and antibody mixture for 2–6 h. Immunocomplexes were washed with NP40, and degenerated with 2×SDS loading buffer including 250 mM Tris-HCl pH 6.8, 10% SDS, 50% glycerinum, 0.5% Bromophenol Blue and 5% β-mercaptoethanol, and subjected to immunoblotting.

HEK293 cells or mCCD cells were cultured on specific confocal dishes. After 5 h, adherent cells were fixed in 5% paraformaldehyde and permeabilized with 5% Triton X-100 in phosphate-buffered saline (PBS) at room temperature. Cells were washed three times with PBS. After blocking with 1% bovine serum albumin for 30 min, cells were incubated with specific antibodies at 4°C overnight. Fluorescent secondary antibodies against rabbit or mouse IgG were incubated for 1 h at room temperature. After washing with PBS, and adding 50% glycerol, the cells were examined by confocal microscopy.

HEK293 cells were cultured in a 3.5 cm cell culture dish. After transfection for 24 h, the cell culture medium was changed to the medium containing 50 μg/ml cycloheximide (freshly diluted from a 100 mg/ml stock in DMSO) and 10 μg/ml Brefeldin A (freshly diluted from a 10 mg/ml stock in ethanol). The cells were lysed at the indicated time points and the cell lysates were analyzed by western blotting.

HEK293 cells were transfected with α-ENaC, and co-transfected with or without HA–derlin-1. Cell lysates were immunoprecipitated with anti-derlin-1 antibody. Protein digestion, labeling, mass spectrometry data acquisition, and identification were completed in the analysis center of Nanjing Medical University. In brief, the labeled peptides were analyzed on a LTQ-Orbitrap instrument (Thermo Fisher) connected to a Nano ACQUITY UPLC system via a nanospray source. The LC-MS/MS was operated in the positive ion model as described previously. The MS/MS spectra acquired from precursor ions were submitted to Maxquant (version 1.2.2.5), and search parameters were followed, as described before (Wang et al., 2013).

Statistical analyses were performed using SPSS 13.0 software. All data are presented as mean±s.e.m. Differences in the various parameters among the groups were evaluated by variance (ANOVA) with post-hoc comparisons. Significance was defined as P<0.05.

We thank Dr Rongfeng Li (Nanjing Medical University) for her critical suggestions and proofreading of this manuscript.